What is Time?
And where is that?
Two years ago I created an introductory programming video course (on the Python Standard Library) in which I needed to document the behavior of Python's datetime module. Obviously, to do that I have to discuss plate tectonics and magma flows, and of course, the tidal locking of orbital bodies (for beginners).
But I think this is an OK characterization of the nature of “time”:
The thing to understand is that a “second” in UTC—and in all the time zones that are defined by offsets from UTC—is not a measure of time. It is a measure of the angular rotation of the Earth. Specifically, it is approximately 15.041 seconds of angular longitude (relative to the sun). This odd number comes from the fact that Earth needs to rotate slightly more than 360° to return to facing the sun because of simultaneous prograde revolution around the sun.
I managed in that course to avoid discussion of the different ground state decay characteristics of the oscillation of rubidium-89 vs cesium-133.
Infuriatingly, the issue of the missing Dec 30, 2011 in Samoa and Tokelau is not addressed in the IANA timezone database, even though it could be. I mean, sure, I get that February 1-13, 1918 in Russia is before the Olson database covers, so that's a pass... but WTF with Samoa?!12
I understand the political events. The problem is getting computer systems to do the right thing. E.g. calculate the duration between 2011-12-29T23:59:00 and 2011-12-31T00:01:00... i. e. two minutes.
My good friend Brad Huntting notes:
No, the UTC second is immutable. It's the UTC minute that's not constant; being unpredictability 60, or 61 seconds or theoretically 58, 59, or 62 seconds.
In light of the future, likely forthcoming 59 second minute, it’s good that he keeps me honest.